Adenomyosis Imaging features for diagnosing the disease and treatment effects of bromocriptine

(1)

From the Department of Women´s and Children´s Health Karolinska Institutet, Stockholm, Sweden

Adenomyosis

Imaging features for diagnosing the disease and treatment effects of bromocriptine

Johanna Andersson

(2)

All previously published papers were reproduced with permission from the publisher.

Published by Karolinska Institutet.

Printed by Universitetsservice US-AB, 2021

Cover illustration: Transvaginal ultrasonographic image from one of the first women that was included in study I, with permission.

(3)

Adenomyosis- Imaging features for diagnosing the disease and treatment effects of bromocriptine

THESIS FOR DOCTORAL DEGREE (Ph.D.)

By

Johanna Andersson

The thesis will be defended in public at conference room Widerströmska 8^th floor, KI Campus, Tomtebodavägen 18a, 2021 May 28 at 9.00 am

Principal Supervisor:

Professor Kristina Gemzell-Danielsson Karolinska Institutet

Department of Women´s and Children´s Health Division of Neonatology, Obstetrics and Gynecology

Co-supervisors:

Associate professor Elisabeth Epstein Karolinska Institutet

Department of Clinical Science and Education Division of Obstetrics and Gynecology

Assistant professor Nageswara Rao Boggavarapu Karolinska Institutet

Department of Women´s and Children´s Health Division of Neonatology, Obstetrics and Gynecology

Associate professor Arne Rådestad Karolinska Institutet

Opponent:

Kirsten Hald, PhD University of Oslo

Department for Research and Development Division of Obstetrics and Gynecology

Examination Board:

Professor Matts Olovsson Uppsala University

Department of Women´s and Childrens Health Division of Obstetrics and Gynecology

Associate professor Greta Edelstam Uppsala University

Department of Women´s and Childrens Health Division of Obstetrics and Gynecology

Associate professor Eva Uustal Linköping University

(4)

(5)

“The glory of medicine is that it is constantly moving forward, that there is always more to learn. The ills of today do not cloud the horizon of tomorrow, but act as a spur to greater effort.”

William Mayo 1926

To all women

(6)

(7)

POPULÄRVETENSKAPLIG SAMMANFATTNING

Första gången sjukdomen nämns är år 1860, då Rokitansky beskriver en växt inuti livmoderväggen innehållande öar av slemhinna. I slutet av 1800-början av 1900-talet beskrevs flertal förändringar i livmoder och bukhinna. Man kallade förändringarna adenomyom, vilket var ett samlingsnamn för de 3 tillstånden muskelknutor, endometrios och adenomyos. Olika teorier framhölls om vad som orsakade fynden och forskare var inte ense. 1920 lyckades Dr Cullen, som var gynekolog, ena forskarna om att särskilja

livmoderslemhinna som återfinns i bukhålan (endometrios) från slemhinna som återfinns i livmoderns muskelvägg (adenomyos) och fastställde att det rör sig om 2 olika sjukdomar.

Dr Cullen beskrev symtomen vid adenomyos som ”långdragen extremt smärtsam menstruationsblödning”. Sjukdomen fick 1925 sitt nuvarande namn, adenomyos.

Trots att det har gått 100 år sedan forskare beskrev adenomyos och endometrios som 2 olika tillstånd, har inte det antagandet levt vidare. Vi tar till oss ny kunskap om endometrios och blir allt bättre på att ta hand om drabbade kvinnor. Forskning om adenomyos hade dock fallit i en nästan 100-årig sömn, för att sakta väckas till liv under de senaste 10-20 åren och ta fart på senare år.

De två tillstånden har gemensamt att de utgörs av livmoderslemhinna på avvikande ställen vilket ger smärta kopplat till menstruationsblödning. Förutom smärtsamma menstruationer har kvinnor med adenomyos även kraftiga blödningar. Många kvinnor lider av båda tillstånden samtidigt, vilket försvårar forskningen. Det är också vanligt att kvinnor med adenomyos även har muskelknutor, vilket ger liknande symtom.

Den snabba utvecklingen av teknik har möjliggjort framsteg inom forskning på kvinnans reproduktionsorgan. Tidigare gjordes så gott som all forskning på livmödrar som tagits bort i samband med kirurgiska ingrepp. Utvecklingen av röntgen såsom magnetkamera och ultraljud har lett till forskning som ej var möjlig tidigare. Nu behöver man inte operera bort en livmoder för att undersöka den, utan kan med radiologiska tekniker följa olika förlopp.

Detta har gjort att kunskapen om olika symtom och diagnoser ökat och därmed möjlighet att behandla tillstånd med läkemedel.

Vi känner inte till den bakomliggande orsaken till varför en del kvinnor drabbas av

adenomyos. Det finns olika teorier. På 90-talet gjordes en del studier på möss, vilket visade

(8)

var att kvinnorna fick mindre blödningsbesvär och mindre mensvärk under de 6 månader som de behandlades med läkemedlet. Kvinnorna genomgick ultraljuds och

magnetkameraundersökning före och efter behandlingen. Vi kunde inte se någon

röntgenologisk skillnad före och efter behandlingen, annat än hos ett fåtal kvinnor. Detta var ett förväntat resultat då 6 månaders behandling sannolikt är för kort tid att förvänta sig röntgenologiska förändringar. För att ytterligare undersöka läkemedlet bromokriptins effekt och verkningsmekanism togs slemhinneprover från livmodern före och efter behandlingen.

Nivåerna av prolaktin i slemhinnan förändrades inte, men gener som styr celldöd ökade och gener som styr tillväxt av slemhinnan hämmades. Detta kan vara en del i förklaringen till varför kvinnorna fick mindre blödningar och mindre smärtor efter behandlingen.

Under studiernas gång blev det tydligt att det saknas klara röntgenologiska kriterier för när en kvinna har adenomyos, både vad gäller undersökning med magnetkamera och ultraljud.

Gemensamma riktlinjer om vilka parametrar som krävs för diagnos saknas och uppfattningen om vad som är av vikt, skiljer sig mellan bedömare. Som patient är det viktigt att få samma diagnos oberoende av vem som har bedömt bilderna. För att undersöka hur samstämmiga olika bedömare är gjorde vi magnetkameraundersökning samt

ultraljudsundersökning på 51 kvinnor med rikliga menstruationer där misstanke om adenomyos fanns. Fyra röntgenläkare bedömde magnetkamerabilderna och fem

gynekologer som är specialiserade i ultraljudsundersökning bedömde ultraljudsbilderna. Vi ville ta reda på hur bedömningarna skildes åt och vilka tecken på sjukdom som flest

bedömare var eniga om. Studien visade att röntgenologerna som bedömde

magnetkamerabilderna var mer eniga i de enskilda tecknen på sjukdom som sågs, jämfört med ultraljudsbedömarna. Trots det var gynekologerna som bedömde ultraljudsbilderna mer eniga i sin slutgiltiga bedömning huruvida kvinnan hade adenomyos eller inte. Studien visar svårigheten att ställa diagnosen, även för erfarna läkare. Resultatet visar att fler

röntgenologiska studier behövs. Kriterier för vad som krävs för att ställa diagnosen behöver fastslås, så att svaren inte skiljer sig beroende på vem som tolkar bilderna.

Den sammantagna slutsatsen av studierna, är att vi som ett första steg behöver fastslå riktlinjer för vad som krävs för att ställa diagnosen adenomyos via magnetröntgen och ultraljud samt öka den kunskapen bland gynekologer. När vi har lärt oss att ställa diagnosen systematiskt kan detta underlätta studier av symtom samt bakomliggande orsaker. Denna kunskap kan därefter leda till riktade läkemedelsbehandlingar mot adenomyos.

(9)

ABSTRACT

Background

Adenomyosis is a benign uterine disease, causing various symptoms including heavy menstrual bleeding (HMB) and pelvic pain. In affected women, endometrial glands and stroma are located in the myometrium surrounded by hypertrophied myometrial tissue. The eutopic endometrium is associated with increased proliferation, high migration and a high invasive capacity. The knowledge of the pathogenesis is largely unknown; however, mice models have shown a link between increased uterine concentration of prolactin (PRL) and the disease. The gold standard in the treatment for hyperprolactinemia is the dopamine agonist bromocriptine. Vaginal administration is effective in reducing serum PRL and has less gastrointestinal side effects than oral administration. Reducing uterine PRL may improve symptoms and could be a possible medical treatment in the future if PRL is associated with the disease. Whether bromocriptine reduce uterine PRL is not known.

In the presence of clinical symptoms adenomyosis diagnosis can be confirmed using Magnetic Resonance Imaging (MRI) or Transvaginal ultrasonography (TVS). The reproducibility has been reported high for both modalities, but consensus criteria for diagnosing the disease are still lacking in both MRI and TVS.

Aim

The overall aim was to examine the effects of the dopamine agonist bromocriptine in women with adenomyosis and to assess agreement between MRI and TVS for imaging features associated with the disease. The specific objectives were to assess symptoms before and after treatment with vaginal bromocriptine. Another objective was to assess changes in MRI and TVS during treatment and to analyze changes in the endometrium regarding protein biomarkers and differentially expressed genes. A further objective was to compare the inter-rater agreement between MRI and TVS for diagnosing adenomyosis and for various features in the same set of women.

Methods and results

In study I, 18 women from Sweden and 1 woman from the USA with regular HMB and suspected adenomyosis were included. Women were treated with a daily dose of 5mg vaginal bromocriptine for 6 months. Self-administered questionnaires were used to assess symptoms at baseline, 3 months, 6 months, and 9 months (3 months after cessation of the

(10)

improvement in quality of life was seen with UFS-QOL. Total EHP and FSFI did not show any significant differences.

Study II was a secondary outcome of study I. The 18 women at the Swedish site underwent MRI and TVS at baseline and after treatment with vaginal bromocriptine. The MRIs were assessed by one radiologist and the TVS were assessed by one gynecologist specialized in gynecologic ultrasound. For MRI, no significant differences were found in Junctional Zone (JZ) max, JZ differential, ratio JZ/myometrium or myometrial cysts. TVS showed a

significant reduction in JZmax and in asymmetric wall thickness. No significant changes were seen in irregular JZ, fan shaped shadowing, striations, hyperechogenic islands, or cystic lesions.

In study III, MRI and TVS images from the same set of 51 women with HMB and suspected adenomyosis were assessed. MRIs were assessed by four radiologists and the TVS images were assessed by five gynecological ultrasonographers. For MRI, the inter- rater reliability for JZ measurements were ‘moderate to good’. Inter-rater agreement for wall asymmetry and irregular JZ were ‘moderate’, while the inter-rater agreement for globular uterus shape was ‘poor’ and ‘fair’ for cysts. The overall subjective impression if adenomyosis was present or not was ‘fair’. For TVS, the inter-rater agreement for globular uterus shape and wall asymmetry were ‘moderate’. Irregular JZ, fan shaped shadowing and buds or striations were ‘fair’ and the inter-rater agreement for cysts were poor.

Measurement of the JZ did not show any agreement between the raters. The overall subjective impression of whether adenomyosis was present or not was ‘moderate’.

Study IV was a secondary outcome of study I. Endometrial biopsies were taken at baseline and after treatment with vaginal bromocriptine. 12 paired (i.e 12x2) samples were included in the study. Analyses were carried out to evaluate PRL and differentially expressed genes before and after treatment were performed. A significant reduction in serum PRL was observed following bromocriptine treatment, but no changes in the eutopic endometrium.

Gene expression analysis showed a significant upregulation of BAX (a marker of apoptosis) and downregulation of Ki67 (a marker of proliferation) and downregulation of genes associated with glucose metabolism.

Conclusion

A significant improvement in menstrual bleeding, pain, and quality of life was seen after 6 months of vaginal bromocriptine treatment. A significant decrease in JZmax and

asymmetric myometrial wall thickness were demonstrated with TVS. An anti-proliferative effect by downregulating genes associated with glucose metabolism was seen in the eutopic endometrium after treatment.

The inter-rater agreement for diagnosis was higher for TVS than for MRI despite MRI manifesting higher agreement in most features associated with the disease.

(11)

LIST OF SCIENTIFIC PAPERS

I. Vaginal bromocriptine improves pain, menstrual bleeding and quality of life in women with adenomyosis

ANDERSSON JOHANNA KARIN, Khan Z, Weaver AL, Vaughan LE, Gemzell-Danielsson K, Stewart EA

Acta Obstet Gynecol Scand. 2019 Oct;98(10):1341-1350

II. Vaginal bromocriptine for treatment of adenomyosis: Impact on magnetic resonance imaging and transvaginal ultrasound

ANDERSSON JOHANNA KARIN, Pozzi Mucelli R, Epstein E, Stewart EA, Gemzell-Danielsson K

Eur J Obstet Gynecol Reprod Biol. 2020 Nov;254:38-43

III. Inter-rater agreement for diagnosing adenomyosis using Magnetic Resonance Imaging and Transvaginal Ultrasonography

ANDERSSON JOHANNA KARIN, Pozzi Mucelli R, Dueholm M, Fridsten S, Grigoriadis A, Guerrierro S, Leone F, Valentin L, Van den Bosch T, Voulgarakis N, Gemzell-Danielsson K, Epstein E

Manuscript

IV. Bromocriptine reduces endometrial cell growth in women with adenomyosis by targeting glucose metabolism

Ponandai-Srinivasan S, ANDERSSON JOHANNA KARIN, Frisendahl C, Korsching E, Tengelin J, Pavone D, Stewart E, Parameswaran Grace Luther L, Rao Bogavarappu N, Gemzell-Danielsson K

Manuscript

(12)

(13)

LIST OF ABBREVIATIONS

2D 3D AMCOQ BMI cDNA COC DE DNA EHP-30 ELISA FSFI GnRH HMB HRQOL HSIL ICC IUD IVF LNG-IUS MD MPQ MRI mRNA OAB PACS

Two Dimensional Three Dimensional

Aberdeen Menorraghia Clinical Outcomes Questionnaire Body Mass Index

Complementary DNA

Combined Oral Contraceptives Differentially Expressed Deoxyribonucleic acid

Endometriosis Health Profile Questionnaire Enzyme-Linked Immunosorbent Assay Female Sexual Function Index

Gonadotropin Realizing Hormone Heavy Menstrual Bleeding

Health Related Quality Of Life

High grade Squamous Intraepithelial Lesion Intraclass Correlation Coefficient

Intrauterine Device In Vitro Fertilization

Levonorgestrel releasing intrauterine system Mean Difference

McGill Pain Questionnaire Magnetic Resonance Imaging Messenger RNA

Overactive Bladder

Picture Archiving and Communication System

(16)

RNA-Seq SD SPSS SSS TCRE TF TVS UAE UFS-QOL VAS

RNA Sequencing Standard Deviation

Statistical Package for the Social science Symptom Severity Subscore

Trans Cervical Resection of the Endometrium Tissue Factor

Trans Vaginal ultrasonography Uterine Artery Embolization

Uterine Fibroid Symptom Quality of Life Questionnaire Visual Analogue Scale

VCI WHO JZ JZdiff JZmin JZmax JZmax/myom

Volume Contrast Imaging World Health Organization Junctional Zone

Junctional Zone differential

Minimal Junctional Zone thickness Maximal Junctional Zone thickness JZmax/myometrium

(17)

1 INTRODUCTION

The Latin word “Adenomyosis” stands for Adeno- glands, myo- muscle, and osis- derangement which describes the disease well. Adenomyosis is a benign disease where glands and stroma from the endometrium are found in the myometrium surrounded by hypertrophic myometrial cells (1, 2).

There are 3 different types of adenomyosis; The most common type is “diffuse”, where the adenomyotic tissue is scattered diffusely throughout the myometrium. In “focal

adenomyosis”, the lesion is located in one restricted area in the myometrium. “Cystic adenomyosis” is rare and consists of hemorrhagic cystic structures.

As far as possible, this thesis focuses on diffuse adenomyosis.

Figure 1. A normal uterus to the left and an uterus with diffuse adenomyosis to the right. The figure is adapted from Dr. Liang, Sydney Fibroid Clinic website

1.1 HISTORICAL BACKGROUND

The first time the disease was mentioned in the literature was in 1860, when Rokitansky described a growth inside the uterus, containing islands of endometrium. The discovery was followed by other similar findings in the uterus and peritoneal cavity. All different findings

(18)

The current definition of adenomyosis was suggested 1972 by the pathologist Bird;

“Adenomyosis may be defined as the benign invasion of the endometrium into the

myometrium, producing a diffusely enlarged uterus which microscopically exhibits ectopic non-neoplastic, endometrial glands and stroma surrounded by the hypertrophic and

hyperplastic myometrium”(1).

It is striking, that we, 50 years later, still have limited knowledge about the disease, and that we are still debating if endometriosis and adenomyosis are one or two different diseases. It is common to be affected by both diseases at the same time and they have similar

symptoms, which makes research in the field challenging.

1.2 PREVALENCE

Adenomyosis is a common disease, but the exact prevalence is not known. In studies conducted over the last 20 years, a prevalence of 23- 43 % (3-7) is reported among women who underwent a hysterectomy. The studies are retrospective and based on histological diagnosis. A limitation of the studies is that only women who need a surgical intervention are included, which may be more severe cases (Figure 2). Other limitations are that both pre- and postmenopausal women are included in most studies and that there are

heterogenous indications for the surgery.

Figure 2: It is important to be aware of the selection bias in retrospective studies including women with a histological diagnose after hysterectomy. The figure is adapted from Upson et al, Semin Reprod Med 2020

(19)

Studies assessing prevalence among women with symptoms associated with the disease are lacking. However, in one study, from Nepal, where there is limited access to preoperative diagnostic tools such as Magnetic Resonance Imaging (MRI) and Transvaginal ultrasound (TVS), 45 % of women that underwent a hysterectomy on the indication heavy menstrual bleeding (HMB), were found to have adenomyosis (7). In studies utilizing TVS in women referred to university clinics, 21-34 % were diagnosed with adenomyosis (8, 9).

Diffuse adenomyosis is the most common form of adenomyosis. One study reported diffuse adenomyosis in 84 % of women with histologically confirmed adenomyosis (4).

Thus, adenomyosis is a common disease but the true prevalence among premenopausal women with HMB and pelvic pain needs to be further explored.

1.3 COMORBIDITY

As many as 80 % (10, 11) of women with adenomyosis and dysfunctional uterine bleeding may have additional pelvic pathology, such as endometriosis or leiomyomas that interfere with the symptoms and makes research in the field challenging.

In women with adenomyosis that underwent a hysterectomy, 40-46 % were reported to have comorbidity with endometriosis (4, 12). Especially deep infiltrating endometriosis seems to be associated with adenomyosis (10, 13, 14).

1.4 MOLECULAR MARKERS

Adenomyosis is suggested to be associated with altered apoptosis and proliferation (15-18).

The protein Ki67 is associated with cell proliferation. The presence of Ki67 was reported higher in the eutopic endometrium in women with adenomyosis than in healthy women (Figure 3) (16), indicating a more rapid proliferation. The same study also showed reduced apoptosis, but not related to the Bcl 2 gene, which is a common marker for apoptosis.

Tissue Factor (TF) is a cytokine that plays a role in angiogenesis and apoptosis. TF has shown to be elevated in both eutopic and ectopic endometrium in women with adenomyosis (18).

(20)

Figure 3: Immunocytochemical staining of Ki67 (brown color)in the endometrium from a woman without adenomyosis (to the left) and a woman with adenomyosis (to the right). The figure is reused from Yang et al 2007, with permission provided by Rights Link Copyright clearance center.

Differentially expressed (DE) genes were seen in adenomyotic mice compared with controls suggesting impaired differentiation of cells during fetal development of the myometrium (19). DE genes in the myometrium in women with adenomyosis compared with myometrium from healthy women are also reported (20). The enzyme aromatase cytochrome P450 catalyzes the conversion of androgens to estrogen and has been shown to be overexpressed in women with adenomyosis and endometriosis (21).

1.5 HISTOLOGY

The disease has, until recently, been a histological diagnosis, even though consensus for histological criterion is lacking. Adenomyosis is histologically defined as endometrial glands invading the basalis layer of the endometrium. Areas of endometrial glands and stroma are found in the myometrium, causing hyperplasia and hypertrophy of the

surrounding myometrium (1). Different minimal depths are suggested (1, 2, 22). However, invasion of at least 2.5 mm below the basalis layer is used in most studies (23, 24). To make the diagnosis, hypertrophy of the surrounding myometrium must be present.

The histological diagnosis is dependent of how the biopsies are sampled and how well they represent the location of the disease. For example, in one study, 31% of the samples

contained adenomyosis when three routine sections where taken, and 61% when six sections where taken (1). There is still no consensus on how the uterus should be sampled after a hysterectomy.

(21)

1.6 PATHOLOGY

The pathology of adenomyosis is unknown, but there are different theories;

1.6.1 Invasion of the endometrium

The most widespread theory proposes that the endometrium grows into the myometrium by invading a traumatized endometrial-myometrial border (25). The trauma is thought to be caused by surgical procedures, such as curettage for termination of a pregnancy or cesarean delivery. An increased incidence of adenomyosis after curettage for termination of a

pregnancy has been reported (6, 26, 27). However, there are also studies where association with curettage or cesarean delivery has not be found (22, 27, 28). Therefore, it has been proposed that it may be the pregnancy itself that causes a disruption of the barrier between the endometrium and the myometrium, not the surgery. Parous women, particularly those with a first birth at an early age, are more likely to develop the disease than nulliparous women (7, 23, 28-32). Further, an increased risk has been shown with an increasing number of births (9, 23).

1.6.2 Müllerian remnants

A commonly proposed theory is that adenomyosis has a Müllerian origin, caused by altered development of the genital tract during embryonic life. The endometrium and the inner myometrium have different origins than the outer myometrium. Displaced embryonic pluripotent Müllerian remnants in the myometrium of the female fetus are thought to be able to differentiate into endometrial cells in the myometrium and cause adenomyosis.

1.6.3 Hyperperistalsis

Hyperperistalsis and increased intrauterine pressure might result in rupture of the basalis layer, leading to infiltration of the endometrium into the myometrium (33-35). Briefly, micro traumatization may lead to a cascade that increases the level of estradiol, which in turn leads to proliferation, angiogenesis and hyperperistalsis. The increased peristalsis would further exacerbate the process with increased injury and ultimately the formation of adenomyotic lesions (Figure 4).

(22)

Figure 4:Model of tissue injury and repair that initiates the genesis of adenomyotic lesions. The hyperperistalsis induces increased production of Prostaglandins, which increase estrogen production. Estrogen activates the estrogen receptor  and , leading to induction of Oxytocin (OT/OTR) that increases uterine contractions and exacerbate the injury. Reused from Guo et al, Journal of Clinical Medicine, with permission.

1.6.4 Immune system

A role of the immune system in adenomyosis has also been considered (36). Proliferation of specific lymphocytes and formation of antibodies may be involved in the tissue damage.

Another reported risk factor are macrophages affecting the endometrial cells and causing subsequent increase in interleukin production (37).

1.6.5 Increased Estradiol exposure

Adenomyosis is shown to be associated with hyperestrogenism (38, 39). Increased estrogen exposure resulting from an early menarche, increasing age, short menstrual cycles and obesity are reported as risk factors (29, 40). Higher concentrations of estradiol in menstrual blood were detected in women with adenomyosis than in women without the disease, whereas the level in serum was equal (41). Furthermore, genes involved in estrogen synthesis are shown to be upregulated in women with adenomyosis (42, 43).

1.6.6 Prolactin

Animal models suggests that adenomyosis is related to increased uterine concentrations of Prolactin (PRL) (44, 45).

PRL is produced in the pituitary gland and in small amounts in the endometrium and the myometrium (46, 47). Prolactin acts as a smooth muscle cell mitogen in vitro (48, 49). The PRL receptor (PRL-R) has shown to be upregulated in adenomyotic uteri of mice and the degree of upregulation correlated to the histologic extent of the disease (50). There is evidence that minimal increase in the level of PRL in serum secondary to medication with

(23)

antidepressants is sufficient to cause adenomyosis both in murine and human uteri (51, 52).

A retrospective human study observed a history of depression as the only factor associated with adenomyosis compared to women with leiomyoma (53).

Reducing the concentration of uterine PRL may be a possible future medical treatment option if PRL is central in the pathogenesis.

1.7 SYMPTOMATOLOGY

1.7.1 Heavy menstrual bleeding

The most common symptom is HMB (29, 32, 54, 55), reported by up to 80% of

symptomatic women (4, 5, 32). The severity of HMB has been shown to correlate with the depth of glandular invasion (27, 56, 57).

1.7.2 Pelvic pain

Dysmenorrhea and other forms of pelvic pain are reported by 30 - 80 % (4-6, 8, 29, 30, 32, 53, 55) of women with adenomyosis. The severity of pain correlates to deeper foci (27, 58).

In one study, where dysmenorrhea was reported by 79 % of women, the mean Visual Analogue Scale (VAS) score was as high as 8 (8). Another study reported that more than half of the women with dysmenorrhea associated with adenomyosis suffered from severe pain (6).

1.7.3 Urinary tract symptoms

A few studies have investigated bladder symptoms and concluded that overactive bladder symptoms (OAB) were significantly more frequent in women with adenomyosis than in healthy controls. Increased daytime frequency of voiding were common and 20-45 % of women with adenomyosis reported urge incontinence (59, 60). The women had TVS findings suggestive of adenomyosis as inclusion criteria. How many of them that had comorbidity with endometriosis were not determined. Thus, whether OAB is a symptom associated with adenomyosis or rather a symptom associated with endometriosis is not known. The underlying cause for OAB in women with adenomyosis needs to be further explored. One reason for the symptom could be pressure on the urinary bladder from an enlarged uterus. Further, chronic pelvic pain leads to nerve sensitization that may cause

(24)

factor for implantation failure (62, 63). Thus, further studies are required to determine the association with adenomyosis and fertility.

1.8 RADIOLOGICAL FINDINGS 1.8.1 Magnetic Resonance Imaging

The ability to make the diagnosis of adenomyosis by MRI has a long tradition and focuses on the inner part of the myometrium, the JZ. Adenomyosis is a disease in the JZ and changes its appearance. However, despite the knowledge of features associated with adenomyosis on MRI, no consensus exists regarding criteria for the diagnosis. The most widespread criteria for diagnosing adenomyosis with MRI is an increased thickness of the JZ ≥12mm (64).

MRI T2-weighted sequences are the key to diagnosing adenomyosis since the sequences highlight the JZ. The myocytes in the JZ have a different morphology from the myocytes in the outer myometrium, with a greater relative nuclear area, a looser extracellular matrix, and lower water content (65). Therefore, on T2-weighted images, the JZ appears darker (hypointense) than the outer myometrium (intermediate intensity) (66). The endometrium appears bright (hyperintense).

There are no guidelines suggesting a MRI protocol for the examination of adenomyosis.

Bowel movements create artifacts that decrease the quality of the image of the uterus.

Therefore, it is an advantage if the women fast before the MRI examination and if antispasmodic drugs are used. To obtain an optimal image of the JZ, an oblique axial sequence perpendicular to the uterine long axis should be taken (11).

The JZ of women with adenomyosis is thicker than in healthy women, due to the smooth muscle hyperplasia that surrounds the ectopic endometrial cells. However, a normal JZ can mimic that of adenomyosis. The JZ is thicker the first days of the menstrual cycle, and contractions can be mistaken for a focal thickening (67). Therefore, adding elements to the criteria for adenomyosis is suggested, such as an irregular appearance of the JZ and its relation to the thickness of the entire myometrium (Figure 5) (11, 68, 69).

The JZ border is often irregular to the myometrium, due to the different depth of the lesion.

The level of irregularity can be measured by the difference between the maximal and minimal thickness of the JZ, the JZ differential (JZdiff). A JZ diff >5 mm may be more reliable for diagnosing adenomyosis than a JZ thickness of >12 mm (11, 68)

Since the JZ is part of the myometrium and adenomyosis is located in the JZ, the percentage of the JZ related to the total myometrium is of interest (JZ/myom). The thickness of the myometrium (healthy myometrium+JZ) is measured at the same level as JZmax. A ratio of 40% indicates adenomyosis (11).

(25)

Intramyometrial cysts are pathognomonic for the disease but are found in only one- third to half of the affected women (11, 64).

1.8.2 Transvaginal ultrasonography

The technique for TVS has developed over the last decade and the ability to visualize signs of adenomyosis has improved since heterogeneity of the myometrium and presence of myometrial cysts were proposed to indicate the disease (24, 64, 68).

Similar to MRI there is no consensus on which diagnosis criteria to include and require for a diagnosis of adenomyosis. Different features on 2-dimensional (2D) ultrasound are associated with the disease (Figure 6). Adenomyosis enlarges the corpus uteri and makes the shape globular (70). It is common that the lesions are more widespread in one of the walls leading to asymmetrical myometrial thickening (8, 71, 72). The endometrial junction appears irregular as the ectopic endometrium grows into the myometrium (70, 71).

Hyperechogenic striations and areas corresponding to ectopic endometrium are found within the myometrium and fan shaped shadowing is present (70).

Myometrial cysts are seen as anechoic areas and are considered highly specific for the disease (70, 73-75).

It is difficult to visualize the JZ with 2D TVS. 3-dimensional (3D) TVS makes it possible

Figure 5A-B: T2-weighted MRIs showing the part of the myometrium (dubbel arrow) that consist of the JZ (line). A) The JZ

>12mm and most of of the myometrium is involved B) The JZ >12mm and less than half of the myometrium is involved.

The images are used with permission from women who participated in the study.

A B

(26)

Figure 6 A-D: TVS images from women in the study. Figure D is reused from paper II. The images are used with permission from the study participants. Different TVS features associated with adenomyosis are shown. A) The uterus has a globular shape and the myometrial walls have asymmetric thickness (dashed dubbel arrow) and hyperechogenic areas in the myometrium are present (thin arrow). B) Fan shaped shadowing is present in the myometrium C) The

endomyometrial junction is irregular and the border is difficult to follow (dashed arrow) D) The endomyometrial junction is irregular and the border is difficult to follow (dashed arrow). Hypoechogenic myometrial cysts (thick arrow) and hyperechogenic endometrial spots are found within the myometrium (thin arrow).

Figure 7: 3D TVS from a woman in the study showing a coronal view of the uterus. The endometrium appears bright. The JZ (dubbel arrow) appears darker than the rest of the myometrium. The image is reused from paper II, with permission.

A B

C D

(27)

1.9 TREATMENT

Traditionally, hysterectomy has been the only treatment option for women with

adenomyosis and is still the only curative treatment. However, uterine sparing surgery is possible and medical treatments are available that reduce HMB and pain in women with adenomyosis.

1.9.1 Uterine sparing surgical interventions 1.9.1.1 Excision by laparoscopy or laparotomy

For diffuse adenomyosis, only part of the lesion will be removed, leaving a defect in the myometrium. Different techniques are used to close the defect in the myometrium. The surgery is only suitable in large centers, due to the high risk of complication.

When resecting diffuse adenomyosis in heavily enlarged uterii, improvement in bleeding and pain has been reported after laparotomy and laparoscopic surgery (76-79). However, most studies have a short follow-up time. Recurrency with regrowth of the adenomyosis is common (78, 79).

1.9.1.2 Trans Cervical Resection of the Endometrium and Ablation techniques

The aim of the treatment is to damage the basal layer of the endometrium to prevent regeneration and thus reduce the amount of menstrual bleeding. Endometrial resection is performed with a loop or rollerball and is operator-dependent while thermal ablation techniques use devices from different companies. Different inclusion criteria are used.

However, the different techniques show similar results. Few studies distinguish women with adenomyosis from other diseases that also cause HMB (leiomyoma, polyps or unknown causes) thus making the evaluation of the results uncertain.

The short-term results for surgery in women with adenomyosis are equal to the results seen in women with other diseases. Improvement in bleeding and pain with a success rate of over 80% is reported (80-83). However, the treatment effect decreases over time and 30%

of women with adenomyosis need further surgery within 3 years (83).

Depth of penetration is an important prognostic factor. Women with deep adenomyosis have an increased risk for treatment failure (80, 82, 84, 85).

1.9.1.3 Uterine artery embolization

(28)

1.9.2 Medical Treatment

Studies focusing on medical treatment for adenomyosis are lacking. Reasons for that may be the comorbidity with leiomyomas and endometriosis together with the fact that

adenomyosis until recently has been a histological diagnosis. There is no available drug that is labeled for adenomyosis. However, most medical treatments for endometriosis and leiomyoma that improve the symptoms HMB and pain may also be effective for adenomyosis.

1.9.2.1 Levonorgestrel intrauterine system and other progestins

The levonorgestrel-releasing system (LNG-IUS), a progestin, is the most studied medical treatment for adenomyosis and the only drug that is labeled for idiopathic HMB. Progestins such as LNG induce downregulation of the estrogen receptor and prevent endometrial proliferation (91) leading to reduced HMB and pain (92). The LNG-IUS releases 20 microgram LNG over 24h. A 90 % reduction in Pictorial Bleeding Assessment Chart (PBLAC) (93, 94) and almost total pain relief (93, 95) at follow-up at 6 months to 5 years has been reported. Other studies report similar results (96, 97). Satisfactory rate is reported to be 70-90% at 3 years after insertion of an LNG-IUS (95, 96). Further, LNG-IUS

improved OAB significantly 6 months after insertion (98).

Notable, in the majority of studies, the results are analyzed in women that continued the treatment for the entire study period. Women who discontinued the study are not included, which will bias the results. The most common reasons for dropping out of the studies are expulsion of the device, continuation of HMB, or pain. Women who are dissatisfied with the treatment frequently chose to remove the device and will thus be excluded from the study, which may give false positive results for PBLAC and VAS in the studies.

Other progestins have high efficacy for reducing pain in endometriosis and are commonly used for adenomyosis even though studies in women with adenomyosis are scarce. In two small pilot studies, Dienogest showed improved pain in women with adenomyosis, but with less improvement in HMB (99, 100).

1.9.2.2 Other hormonal treatments

Combined oral contraceptives (COCs) with estrogen and progestin induce decidualization and subsequent atrophy of the endometrium and reduces HMB and pain (101). COCs is commonly used for the treatment of adenomyosis with satisfactory pain and bleeding control but reported to be less effective than LNG-IUS (102).

Gonadotropin realizing hormone (GnRH) suppresses ovarian estrogen production. GnRH analogues have shown to be effective in the treatment of HMB and pain in women with adenomyosis (100, 103). GnRH also decreases the expression of aromatase P450 (104).

Aromatase cytochrome P450, the enzyme that catalyzes the conversion of androgens to

(29)

estrogen, is overexpressed in women with adenomyosis (21). Thus, besides causing a hypoestrogenism in the endometrium, GnRH may have a specific effect on the disease.

Aromatase inhibitors are in clinical use for treating adenomyosis and are believed to be effective. However, supporting evidence from the scientific literature is scarce (105).

1.10 BROMOCRIPTINE

Bromocriptine is a dopamine receptor agonist that inhibits pituitary secretion of PRL. It is approved since 1978 and is the gold standard of treatment for hyperprolactinemia. Other indications include Parkinson´s disease, acromegaly and inhibition of lactation in

breastfeeding women. In the USA the drug is also approved for the treatment of type 2 diabetes (106).

The most common side effects of bromocriptine are nausea, headache, tiredness, dizziness and vomiting . Uncommon side effects include low or high blood pressure, orthostasis, fainting, compulsive behavior such as an increased urge to spend money, gambling and a hypersexual behavior. Serious side effects are rare but include myocardial infarction, arrhythmia, pericarditis and changes in the cardiac valve. Contraindications includes hypersensitivity to ergot alkaloids, syncopial migraines and uncontrolled hypertension.

Vaginal administration is well tolerated and effective in reducing circulating PRL levels in women with hyperprolactinemia (107, 108) and has fewer gastrointestinal side effects than oral administration (107, 109).

Bromocriptine increases glucose tolerance and decreases insulin resistance, leading to improved glycemic control (110, 111). The underlying mechanism of action is not known in detail.

1.11 SUMMARY

Comorbidity with leiomyoma and endometriosis is very common. Until recently,

adenomyosis has been a histological diagnosis only possible after a hysterectomy. Taken together this may have contributed to the lack of knowledge about the disease. Still, the

(30)

(31)

2 RESEARCH AIMS

The overall aims of this thesis was to examine the effects of the dopamine agonist

bromocriptine in women with adenomyosis and to assess agreement between MRI and TVS for diagnosing the disease.

The specific aims of the different studies were:

Study I: To assess subject symptoms (bleeding and pain) and evaluate quality of life, in women diagnosed with adenomyosis before and after treatment with vaginal bromocriptine.

Study II: To assess changes in the uterus by MRI and TVS in women with adenomyosis before and after treatment with vaginal bromocriptine.

Study III: To compare the inter-rater agreement between MRI and TVS for diagnosing adenomyosis and for various features in the same set of women.

Study IV: To analyze endometrial biopsies from patients with adenomyosis before and after bromocriptine treatment, for RNA and protein biomarkers and to decipher the mechanism of action for bromocriptine.

(32)

(33)

3 MATERIAL AND METHODS

3.1 TABULATED OVERVIEW OF STUDIES

Design and study subjects

Outcome Subjects in final analyses

Method Statistical

analyses

Study I Pilot study including women with adenomyosis treated with vaginal bromocriptine for 6 months.

Changes in bleeding, pelvic pain, and quality of life after treatment.

19 Study visit 1. Baseline

2. 3 months of treatment 3. 6 months of treatment

4. 9 months from baseline. 3months after cessation of study drug.

Subjects filled in PBLAC, VAS, MPQ, UFS-QOL, EHP30, and FSFI each visit.

Descriptive statistics

Study II Secondary outcome from study I. MRI and TVS images before and after treatment with bromocriptine.

Changes in radiological features associated with adenomyosis after treatment.

18 1 radiologist evaluated the MRI images and 1 gynecologic

ultrasonographer evaluated the TVS images from the women before and after treatment.

Study III Reproducibility study including MRI and TVS images from women with HMB and suspect adenomyosis.

Agreement between raters evaluating MRI images compared with raters evaluating TVS images.

51 4 radiologists evaluated the MRI pictures and 5 gynecologic ultrasonographers evaluated the TVS images.

Inter-rater agreement, Fleiss kappa

Interclass correlation, ICC

Study IV Secondary outcome from study I.

Endometrial Pipelle biopsies from study subjects in study I.

Alterations of genes following bromocriptine treatment.

12 Bioinformatic analyses PCR

Immunohistochemistry ELISA

3.2 STUDY DESIGN AND STUDY SUBJECTS. STUDY I, II AND IV

This thesis is mainly based on data from a prospective single-arm pilot study conducted in a collaboration between the Mayo Clinic, Rochester, USA, and Karolinska Institutet.

The protocol for study I was designed by the research team at the Mayo Clinic where the

(34)

3.2.1 Tabulated study protocol

Enrollment Baseline Variable time to reach 5mg bromocriptine

1 month

3 month

6 month

9 month

Phone call X X

Visit X X X X X

Blood Sample

X X

MRI X X

TVS X X

Informed consent

X Baseline

characteristic questionnaire

X

PBLAC^# X X X X

AMCOQ^# X X X X

VAS^# X X X X

MPQ^# X X X X

UFS-QOL^# X X X X

EHP 30^# X X X X

FSFI^# X X X X

Endometrial biopsy

X X

# Abbreviation of questionnaire explained in Material and Method section

3.2.2 Study subjects

The study subjects at the Swedish site were recruited between January 2014 to April 2016.

54 women were screened for enrollment in Sweden, of which 34 were eligible, and 22 were included. Four women dropped out and 18 women completed the study in Sweden. In addition, one woman was recruited and completed the study in the USA. This, a total of 19 women completed the study and were included in the analyses.

Women 35-50 years with regular HMB, MRI or TVS highly suggestive of adenomyosis and not using antidepressants, hormonal contraceptives, or an intrauterine device (IUD), were recruited. Further, women with known or suspected endometriosis, leiomyomas

>4cm, or multiple leiomyomas, were not considered for the study.

3.2.2.1 Baseline characteristics

Women enrolled in the study were 44.8±3.5 years, with an average body mass index (BMI) of 26.5±3.9. 84 % were parous and 63 % had no prior cesarean deliveries . At baseline, all women reported HMB, 78 % reported PBLAC > 250. 79 % used >10 tampons or pads on their heaviest day of menstrual bleeding, 79 % reported menses lasting >7 days and 68 % reported moderate to severe cramps with menses. The history of HMB and pain was long-

(35)

standing; 68 % reported painful menses and 47 % reported HMB before the age of 20.

While few women had a history of blood transfusion (5 %), the majority had previously been on pharmacotherapy for anemia (63 %). The use of medication for painful periods was reported by 63 % of the women.

3.3 STUDY DESIGN AND STUDY SUBJECTS STUDY III 3.3.1 Study design

The study was a reproducibility study comparing agreement among raters of MRI and TVS images. Images from 51 women examined from January 2014 to December 2016 were collected.

3.3.2 Study subjects

Women aged 35-50 years with regular HMB and suspected adenomyosis on clinical examination (including TVS) were referred to MRI and expert TVS as part of the

investigation. All women were clinically examined and recruited by the doctoral student.

Women with adenomyoma, known endometriosis, uterine leiomyomas >4cm, multiple leiomyomas, or current use of an IUD or hormonal contraception were not eligible for the study.

In total 67 women were referred to MRI and TVS. MRI was missing in 3 women (declined examination (n=2); wrong identification number (n=1)) and TVS was missing in 13 women (images were not pseudonymized (n=2); 3D volumes (n=6) or 2D video sequences (n=5) were not recorded or of insufficient quality). Both MRI and TVS images were available in 51 women and included in the study.

3.4 METHODS STUDY I

Women meeting the inclusion criteria and no exclusion criteria were included in the study.

Women with suspect adenomyosis were referred to the department of diagnostic imaging, Aleris Specialist care Sabbatsberg for MRI and to Karolinska University Hospital for expert TVS examination before inclusion.

(36)

headache). The protocol was therefore modified to a slower dosage increase. The women were told to start with ½ tablet (1,25mg) once daily for 1 week and then increasing with ½- 1tablet a week until reaching the final dose of 5mg per day. After the modification of the protocol, no further women dropped out due to side effects.

The enrolled women received a phone call from the study coordinator every week during the first month, to check for compliance, side effects and decide when to increase the dosage. The first day the woman successfully took 5 mg bromocriptine was considered study day 1. The medication was then continued for 6 months with visits in the proliferative phase of the menstrual cycle after 3 and 6 months. At the 3 months visit, 200 tablets of bromocriptine were distributed. The 9 months visit was scheduled 3 months after cessation of the study drug. Subjects were instructed to complete the questionnaires close to their last day of menstrual bleeding during the proliferative phase of the cycle, at baseline and at 3, 6, and 9 months. Women were provided with tampons and sanitary pads for standardization of PBLAC reporting. Women were instructed to refrain from the use of nonsteroidal anti- inflammatory drugs or tranexamic acid therapy during the periods when bleeding was assessed (baseline, 3, 6, and 9 months) to avoid confounding. Scores obtained from these questionnaires were compared between baseline and, 3, 6, and 9 months.

3.4.1 MRI

Women were considered to have adenomyosis if on MRI the JZ ≥12 mm, JZ diff >5 mm and JZ/myom >40 % (11, 68). The presence of cystic changes was also considered.

3.4.2 TVS

Women were considered to have adenomyosis through pattern recognition; Prescence of globular uterus shape, asymmetric thickness of the myometrial walls, irregular endometrial myometrial junction, fan shaped shadowing in the myometrium and myometrial cysts.

3.4.3 Questionnaires

In 2014, when the study started, there were no quality of life questionnaires validated for adenomyosis or translated to the Swedish language. To measure blood loss, pain, and quality of life, a variety of questionnaires were used that are validated and frequently used in studies for diseases with similar symptoms, namely endometriosis, and leiomyomas. The questionnaires were translated from English to Swedish by the doctoral student and the main supervisor. The questionnaires were distributed to the women in advance, filled in at home by the woman at the end of the menstrual period, and handed in on study visits. The questionnaires were sent to the Mayo Clinic, Rochester, USA, and transferred into the Research Electronic Data Capture (REDCap^®, Vanderbilt University) data entry and management program (112, 113) by statisticians.

(37)

3.4.3.1 Baseline characteristics questionnaire

Questions included baseline characteristics such as age, BMI, medical history, smoking, heredity, and physical habits. The questionnaire also consisted of gynecologic history and related questions such as gravidity, parity, cesarean section, fertility problems, use of contraceptives, age when the HMB and pain began, previous medication, and when the woman got the diagnosis of adenomyosis.

3.4.3.2 Pictorial Blood Loss Assessment Chart (PBLAC)

PBLAC is a subjective assessment of the volume of blood loss during each menstrual period, based on the degree of soiling of sanitary pads and tampons (Figure 8) (114). The degree of soiling is scored 1-20. Total scores >100 are considered HMB. PBLAC can be used as an evaluation tool for treatment outcomes (115).

Figure 8: PBLAC from one of the women that participated in study I.

3.4.3.3 Uterine Fibroid Symptom Quality of Life (UFS-QOL)

The UFS-QOL is a questionnaire that assesses symptom severity (SSS) and health-related quality of life (HRQOL) in women with leiomyomas over the preceding 3 months (116).

The UFS-QOL has become a standard instrument in the USA. It is composed of 2 parts, the

Pad Tampon Leakage

small large

Score

1 point 5 points 20 points 1 point 5 points 10 points 1 point 2 points 1 point

1 IIIII

2 IIIIII

3 IIIII IIIII II

4 IIIII IIIII IIII III IIIII II

5 IIIII IIIIII III III IIIII III

6

7 Tot: 662

8

Clots

(38)

3.4.3.4 Aberdeen Menorrhagia Clinical Outcomes Questionnaire (AMCOQ)

The AMCOQ is an instrument measuring blood loss, specifically for women with HMB (119, 120). It is comprised of 13 questions to assess the amount of bleeding and impact on daily living over the previous three months. The total score ranges from 0-100, where a higher score indicates worse symptoms.

3.4.3.5 McGill Pain Questionnaire (MPQ) and Visual Analogue Scale (VAS)

Pain was assessed by MPQ (121) and VAS. We used the short form of MPQ that uses 15 words to describe three attributes of present pain (affective, evaluative and sensory) (122).

The subjects assess how well the words describes the present pain the last week. The words

“last week” were changed to “last menstruation” in this study. Each selected word was scored 0 (none) to 3 (severe). The total score is obtained by summing the item scores (range 0-45). A higher MPQ score indicates more severe pain.

VAS is a standard scale for pain assessment (scale 0-10) (123). A higher VAS score indicates more severe pain. The VAS scale is integrated into the MPQ-questionnaire as the last question “ How do you evaluate the pain the worst day of your last menstruation?”

3.4.3.6 Endometriosis Health Profile (EHP30)

The EHP30 has become a standard instrument to assess symptoms of endometriosis over the preceding four weeks (124). EHP30 consists of two parts, the “core” questions and the

“modular” questions. The core part consists of 30 questions in five core measures (pain, control and powerlessness, emotional well-being, social support and self-image), and the modular part consists of 23 questions (work, sexuality, fertility, relation with children, relation with medical profession and medical treatment). It is possible to answer “not applicable” in the modular part. The word “endometriosis” in the instruction was changed to “adenomyosis” in this study. Higher scores in the core questionnaire indicate severe symptoms.

3.4.3.7 Female Sexual Function Index

The FSFI is an instrument measuring sexual function in women over the preceding four weeks (125, 126). It consists of 19 questions in six relevant domains using a 5-point likert scale. Scores ranging from 2-36. A higher score indicates better sexual function.

3.5 METHODS STUDY II

Study II is a secondary outcome from study I. The 18 women from the Swedish site who were included in study I underwent MRI and TVS at baseline and after 6 months of treatment with bromocriptine. The images were assessed for evaluating differences in features before and after treatment.

(39)

3.5.1 MRI

MRI of the pelvis was performed on a 1,5T system (Optima MR450w, GE Healthcare, Waukesha, WI, USA or Siemens Magnetom Symphony Tim, Siemens-Healthineers, Erlangen, Germany). The minimum acquired protocol included the following sequences: a T2-weighted Fast Relaxation Fast Spin Echo or a Turbo Spin Echo on the axial, sagittal and coronal plane (slice thickness 4-5 mm; gap: 10-20 %); a T1-weighted Fast Spin Echo or a Gradient Echo on the axial and coronal plane (slice thickness 5 mm; gap 10-20 %). All the examinations were performed with a phase array coil. The participants were asked to fast four hours before the examination. No antispasmodic drugs were administrated to the patients.

All MRIs were evaluated on a Picture Archiving and Communication System (PACS) (Sectra AB, Linköping, Sweden), by the same dedicated radiologist working at the

Karolinska University Hospital. The following parameters were recorded on T2-weighted images: length of the uterus excluding the cervix, antero-posterior and latero-lateral diameters; measurement of the thinner and thicker uterine wall excluding leiomyomas;

JZmax and JZmin measured on a midsagittal image through the long uterine axis; The JZdiff correspond to the calculated difference between JZmax and JZmin. JZmax/myom is calculated by dividing the JZmax and the corresponding myometrial thickness measured at the same level. Further, the presence of cystic changes in the JZ (defined as foci of high signal intensity on T2-weighted and/or on fat-saturated T1-weighted images) was also recorded.

3.5.2 TVS

TVS was performed using a high-end ultrasound system Voluson E10, GE Healthcare (GE Medical Systems, Zipf, Austria) with a 5-9 MHz transvaginal probe. 2D grayscale volumes and video clips comprising the whole uterine body were collected. The software 4D View GE Healthcare (GE Medical Systems, Zipf, Austria) was used to process 3D volumes.

All examinations were performed by a single gynecologist specialized in gynecologic ultrasonography. The women were examined in the lithotomy position with an empty bladder. The uterus was scanned in the sagittal plane from cornu to cornu and in the transverse plane from the cervix to the fundus. The following parameters were recorded:

anteroposterior diameter of the uterus; uterine length, excluding cervix; uterine anterior and posterior wall thickness, excluding leiomyomas, presented as the thickest and thinnest wall,

(40)

3.6 METHODS STUDY III

MRI and TVS images from the same set of 51 women were assessed. MRIs were assessed by four experienced radiologists working at the Karolinska University Hospital, Stockholm.

TVS were assessed by five experienced gynecological ultrasonographers from different centers in Europe. Assessment of the images was performed between December 2019 and February 2020. The raters were blinded to the clinical history, physical examination, and the evaluation of MRI and TVS images made by other raters. Each rater entered the

assessments into the REDCap data entry and management program (112, 113) hosted at the Karolinska Institutet. The raters could save their assessments and resume later, to reduce the risk of fatigue.

3.6.1 MRI

MRI of the pelvis was performed on a 1.5T system (Optima MR450w, GE Healthcare, Waukesha, WI, USA or Siemens Magnetom Symphony Tim, Siemens Healthineers, Erlangen, Germany). The minimum protocol included the following sequences: T2- weighted Fast Relaxation Fast Spin Echo or Turbo Spin Echo in the axial, sagittal and coronal plane (slice thickness 4-5 mm; gap: 10-20 %); T1-weighted Fast Spin Echo or a Gradient Echo in the axial and coronal plane (slice thickness 5 mm; gap 10-20 %). All examinations were performed with a phased array coil. The women were asked to fast for 4 hours before the examination. Antispasmodic drugs were not administered.

The images were pseudonymized and evaluated on a PACS at the Karolinska University Hospital by four experienced radiologists. With regard to the presence or absence of adenomyosis, each rater based the assessment on their subjective evaluation of the

radiological features. With regard to the presence or absence of adenomyosis, there were no standardized criteria given on when to make the diagnosis. The predetermined features assessed are listed in Table 1.

3.6.2 TVS

All women underwent ultrasound examination by a single expert examiner using a high-end ultrasound system Voluson E10 or E8, GE Healthcare (GE Medical Systems, Zipf, Austria) with a 5-9 MHz 3D transvaginal probe. 2D grayscale volumes and video clips, and 3D-VCI volumes including the whole uterine body were saved. The GE 4D View software (GE Healthcare, Wood Dale, IL, USA) was used to assess the 3D-VCI volumes. The 2D volumes and video clips and 3D volumes for each case were pseudonymized and

downloaded to memory sticks and sent to five experienced ultrasonographers. The raters used their own personal computers to assess the volumes. The volumes could be modified during the analysis to optimize the assessment (remove the VCI function, change slice thickness or grey mix, rotate the volume in any plane). With regard to the presence or absence of adenomyosis, there were no standardized criteria given on when to make the

(41)

diagnosis. The JZ was assessed at the coronal plane in the 3D volumes. The predetermined TVS features assessed are listed in Table 1.

Table 1. Predetermined features that were assessed by MRI and TVS and entered into REDCap

Feature MRI TVS

Globular uterus shape yes/no yes/no

Asymmetric wall thickness yes/no yes/no Irregular/interrupted JZ yes/no yes/no Cysts in the JZ/myometrium yes/no yes/no

Fan shaped shadowing - yes/no

Buds or striations - yes/no

Hyperechogenic Islands - yes/no

JZ max In millimeter In millimeter

JZ min In millimeter In millimeter

Myometrial thickness at the same level as JZmax

In millimeter - Ratio JZmax/Myometrium Automatically

retrieved by REDCap

-

JZ differential (JZmax- JZmin)

Automatically retrieved by REDCap

-

Diagnosis: Does the subject have adenomyosis?

yes/no yes/no

3.7 METHODS STUDY IV 3.7.1 Sample collection

Study IV is a secondary outcome from study I, analyzing endometrial biopsies from women participating in the clinical trial. Endometrial Pipelle biopsies were taken in the

proliferative phase of the menstrual cycle at baseline and after 6 months of treatment with bromocriptine. The baseline and 6 months visits were scheduled as follows: The women sent an email or phone call to the study coordinator on day 1 of the menstruation. An appointment was scheduled for the first possible day after the last day of bleeding. The

(42)

poor RNA quality or unavailability of paired biopsies. Therefore, a total of 12 paired endometrial samples, of which 3 were classified as “good responders”, were included in the study. Transcriptomic analysis to evaluate DE genes and small RNAs before and after bromocriptine treatment was performed. The protocol is briefly described in this section.

Details can be found in paper IV.

3.7.1.1 Library preparation

Total RNA was extracted from tissue stored in RNA later using RNeasy total RNA Kit (Qiagen, Hilden, Germany). The quantity and quality of purified RNA was measured using the Qubit high sensitivity RNA assay kit with Qubit 4 Fluorometer (Invitrogen, Singapore) and RNA high sensitivity 6000 Pico Kit on bioanalyzer respectively (Agilent Technologies, Santa Clara, US). cDNA library preparation for next generation mRNA sequencing was performed using the SMART-seq2 protocol (127). Tagmentation of DNA and addition of sample index barcodes were performed. The final amplified libraries were cleaned and the samples were pooled and sequenced. cDNA libraries for small RNA sequencing were constructed according to a highly sensitive small RNA sequencing protocol (128). The small RNA libraries were indexed and the samples were pooled and sequenced. Both mRNA and small RNA sequencing were performed on an Illumina Nextseq 550 sequencer at the Bioinformatics and Expression Analysis core facility at Karolinska University Hospital, Sweden.

3.7.1.2 Bioinformatic analysis

mRNA sequencing data were analyzed for DE genes between the paired samples using Partek flow genomic analysis software (Partek, St. Louis, USA). Filtered DE genes were explored for treatment effect using hierarchical clustering and pathway enrichment tools.

Small RNA sequencing data were analyzed for DE genes between the paired samples using Bioconductor software DESeq (version 1.24.0). Raw and processed data files from both mRNA and small RNA sequencing were deposited in NCBI’s Gene Expression Omnibus.

3.7.1.3 Gene expression analysis

Extracted RNA from the paired samples were converted to cDNA using SuperScript^® VILO^TM kit (Invitrogen®, Thermo Fisher Scientific, Waltham, USA). The gene expression patterns of the top three downregulated glycolytic genes were validated using the following set of Taqman® gene probes: PGK1, GAPDH, ENO1, ki67, BAX, and PRL. cDNA was used in the real-time PCR and analyzed on a Step One Plus Real-time PCR system (Applied Biosystems, USA).

3.7.1.4 Immunohistochemistry to detect Ki67

Immunohistochemistry was performed on paraffin-embedded endometrial tissues using a standardized protocol (129). Primary antibody against Ki67 was diluted using diluent DaVinci Green (Biocare Medical, Concord, CA) and incubated overnight at 4 °C. Rabbit

Adenomyosis Imaging features for diagnosing the disease and treatment effects of bromocriptine

From the Department of Women´s and Children´s Health Karolinska Institutet, Stockholm, Sweden

Adenomyosis

Imaging features for diagnosing the disease and treatment effects of bromocriptine

Johanna Andersson

Adenomyosis- Imaging features for diagnosing the disease and treatment effects of bromocriptine

THESIS FOR DOCTORAL DEGREE (Ph.D.)

Johanna Andersson

POPULÄRVETENSKAPLIG SAMMANFATTNING

ABSTRACT

LIST OF SCIENTIFIC PAPERS

CONTENTS

LIST OF ABBREVIATIONS

1 INTRODUCTION

A B

A B

C D

2 RESEARCH AIMS

3 MATERIAL AND METHODS